Openshift: User-pods not to be scheduled on Dedicated Infrastructure nodes

[Sarath]

Version: Openshift-3.9 && Contrail 5-0-214

As per Openshift core specs, Dedicated Infrastructure nodes not to get User pods scheduled,
https://docs.openshift.com/container-platform/3.9/install_config/install/advanced_install.html#configuring-node-host-labels

[root@server1 ~]# oc get nodes --show-labels
NAME STATUS ROLES AGE VERSION LABELS
server1 NotReady compute,master 15h v1.9.1+a0ce1bc657 beta.kubernetes.io/arch=amd64,beta.kubernetes.io/os=linux,kubernetes.io/hostname=server1,node-role.kubernetes.io/compute=true,node-role.kubernetes.io/master=true,opencontrail.org/controller=true
server2 Ready compute 15h v1.9.1+a0ce1bc657 beta.kubernetes.io/arch=amd64,beta.kubernetes.io/os=linux,kubernetes.io/hostname=server2,node-role.kubernetes.io/compute=true
server3 Ready compute 15h v1.9.1+a0ce1bc657 beta.kubernetes.io/arch=amd64,beta.kubernetes.io/os=linux,kubernetes.io/hostname=server3,node-role.kubernetes.io/compute=true,region=infra
[root@server1 ~]#

[root@server1 ~]# oc get pods --all-namespaces -o wide
NAMESPACE NAME READY STATUS RESTARTS AGE IP NODE
ctest-namespace-20639068 ctest-nginx-pod-43011451 1/1 Running 0 1h 10.47.255.250 server3
ctest-namespace-20639068 ctest-nginx-pod-48813050 1/1 Running 0 1h 10.47.255.251 server3
ctest-namespace-41434280 ctest-busybox-pod-78136594 1/1 Running 0 1h 10.47.255.249 server2
kube-system config-zookeeper-948wf 1/1 Running 0 14h 10.0.0.6 server1
kube-system contrail-agent-2v6t7 2/2 Running 1 14h 10.0.0.4 server2
kube-system contrail-agent-q5xk2 2/2 Running 1 14h 10.0.0.5 server3
kube-system contrail-analytics-jtvtj 7/7 Running 1 14h 10.0.0.6 server1
kube-system contrail-analytics-zookeeper-4grwv 1/1 Running 0 14h 10.0.0.6 server1
kube-system contrail-analyticsdb-vpjgd 2/2 Running 0 14h 10.0.0.6 server1
kube-system contrail-configdb-sxcvv 2/2 Running 0 14h 10.0.0.6 server1
kube-system contrail-controller-config-2dqmn 5/5 Running 0 14h 10.0.0.6 server1
kube-system contrail-controller-control-8xp5h 4/4 Running 0 14h 10.0.0.6 server1
kube-system contrail-controller-webui-726ll 2/2 Running 0 14h 10.0.0.6 server1
kube-system contrail-kube-manager-fkck5 1/1 Running 0 14h 10.0.0.6 server1
kube-system kafka-lc4cw 1/1 Running 0 14h 10.0.0.6 server1
kube-system rabbitmq-9g46f 1/1 Running 0 14h 10.0.0.6 server1
kube-system redis-rxhk5 1/1 Running 0 14h 10.0.0.6 server1
openshift-web-console webconsole-d75b95676-2kwwv 1/1 Running 0 14h 10.47.255.252 server3
[root@server1 ~]#

[Pragash Vijayaragavan (pvijayaragav)]

no testcases fail due to this issue, will be removing sanityblocker

[Pragash Vijayaragavan (pvijayaragav)]

user has to use the command
"oc adm manage-node --selector=region=infra --schedulable=false"
to make this happen. We cannot support this in our playbooks or inventory due to openshift ansible structure.

[Sarath (nsarath)]

Pragash, as this is important recommendatation from Openshift guides, so please make it on TechPub and wiki etc before close this bug as "Fixed" to reflect Techpub changes.